TRANSACTIONS OF THE SECTIONS. 21 



the semidiurnal law of atmospheric pressure is the same at all heights up to 6200 

 feet (on a sharp peak), from 9 p.m. to 9 a.m., both as regards epoch and range. The 

 day variation (9 a.m. to 9 p.m.) is greatest for the lowest station, depending evidently 

 on the temperature. The author connected these facts with the hypothesis proposed 

 by Dr. Lamont and himself, that the semidiurnal variation is due to the inducing 

 electrical action of the sun on our earth and its atmosphere. 



These and several other results, at present only partly worked out, would be pub- 

 lished soon in detail. The printing of the observations made in the observatory of 

 His Highness the Rajah of Travancore was proceeding as rapidly at Trevandrum as 

 could be expected in Jndia, and the first sheets were in the author's hands. 



On the Diurnal Variations of the Magnetic Declination at the Magnetic 

 Equator, and the Decennial Period. By John Allan Broun, F.R.S. 



Solar-diurnal Variation. — The author stated that the observations made at the 

 intertropical observatories had shown the fact that the law of solar-diurnal variation 

 was opposite, or nearly opposite, at two seasons of the year ; this result was made 

 generally known to the scientific world by General Sabine, in his discussion of the 

 St. Helena Observations. St. Helena, however, is too far from the magnetic equator 

 to show the change from one law to the other, otherwise than as a shifting move- 

 ment of the maxima and minima, which seem to slide in the course of a month or 

 two from one position to the other ; the whole range of the variation being consider- 

 able at all seasons. Mr. Broun offered to the Section the results of five years' 

 observations made at the Trevandrum Observatory, about 90 miles south of the 

 magnetic equator, which showed perfectly the mode of variation of the diurnal law. 



In the months of December, January, and February, the minimum of easterly de- 

 clination occurs at 7i a.m., in the months from April to September, the maximum 

 occurs at exactly the same time. In the months of March and October a period of 

 indifference is attained, when the variation becomes nearly zero, or the variation is 

 a series of maxima and minima at different hours, and the range or total angular 

 movement is reduced to about thirty seconds (0 ,- 5) when the mean of a few days 

 is taken. The epochs at which this change takes place are neither those of the sun's 

 crossing the equator nor the zenith, and the epoch seems to vary from year to year. 

 The March epoch is not distant from the vernal equinox, but the other occurs nearly 

 a month later than the autumnal equinox. So far is the second epoch for the change 

 from one law to the other from that of the sun's crossing the zenith or equator, that 

 August and September are the months of greatest diurnal range. 



Although Trevandrum is in 8£ s north latitude, it has a magnetic dip of 2£° south ; 

 but the diurnal variations affect the character of the northern hemisphere more than 

 that of the southern hemisphere, — the mean range for the months from May to 

 September being nearly three minutes (3'), while for the months of December' and 

 January the range is only about two minutes (2'). 



Mr. Broun was the first to point out that the diurnal law at any place might be 

 represented by the superposition of two variations, one resembling that peculiar to 

 high north latitudes, the other resembling that peculiar to high south,— the northern 

 part being always in excess in high magnetic north latitudes, and in excess for places 

 in low magnetic north latitudes only for the sun north of a given line. It is evident 

 that we may, by descending towards the magnetic equator, reach a station where for 

 a given position of the sun the two variations will be equal or nearly equal, in which 

 case for that position of the sun the complete extinction of the diurnal law may be 

 expected ; this occurs approximately at Trevandrum. 



When the diurnal range is examined with reference to the decennial period, it is 

 found that the mean range had a minimum in the year 1856, the exact epoch of 

 minimum being, perhaps, about January of that year ; but when the ranges for 

 given months are considered, some curious differences from the law are discovered. 

 The yearly mean of monthly mean ranges, with the mean ranges for the months of 

 March and October, are as follows : — 



